import tensorflow as tf
from tensorflow.keras.preprocessing.image import ImageDataGenerator
from tensorflow.keras.applications import EfficientNetB0, MobileNetV2
from tensorflow.keras import layers, models
import matplotlib.pyplot as plt
import numpy as np
import osBASE_DIR = "G:\Mi unidad\Colab Notebooks\AProfundo\clasificacion_dataset/"
IMG_SIZE = (224, 224)
BATCH = 32<>:1: SyntaxWarning: invalid escape sequence '\M'
<>:1: SyntaxWarning: invalid escape sequence '\M'
C:\Users\Dieg0AkD\AppData\Local\Temp\ipykernel_21372\1521194072.py:1: SyntaxWarning: invalid escape sequence '\M'
BASE_DIR = "G:\Mi unidad\Colab Notebooks\AProfundo\clasificacion_dataset/"train_gen = ImageDataGenerator(rescale=1/255)
val_gen = ImageDataGenerator(rescale=1/255)
train_ds = train_gen.flow_from_directory(
BASE_DIR+"train",
target_size=IMG_SIZE,
batch_size=BATCH,
class_mode="categorical"
)
val_ds = val_gen.flow_from_directory(
BASE_DIR+"val",
target_size=IMG_SIZE,
batch_size=BATCH,
class_mode="categorical"
)
num_classes = len(train_ds.class_indices)Found 234 images belonging to 3 classes.
Found 47 images belonging to 3 classes.import matplotlib.pyplot as plt
# Mostrar 5 imágenes del dataset de entrenamiento
def mostrar_5_iniciales(dataset):
images, labels = next(dataset) # primer batch
plt.figure(figsize=(12, 8))
for i in range(5):
plt.subplot(1, 5, i+1)
plt.imshow(images[i])
plt.axis("off")
plt.suptitle("Primeras 5 imágenes del dataset (train)")
plt.show()
mostrar_5_iniciales(train_ds)
efficient_model = EfficientNetB0(
include_top=False,
input_shape=IMG_SIZE+(3,),
weights="imagenet"
)
efficient_model.trainable = False
model_e = models.Sequential([
efficient_model,
layers.GlobalAveragePooling2D(),
layers.Dense(256, activation="relu"),
layers.Dense(num_classes, activation="softmax")
])
model_e.compile(
optimizer="adam",
loss="categorical_crossentropy",
metrics=["accuracy"]
)
history_e = model_e.fit(train_ds, validation_data=val_ds, epochs=10)Epoch 1/10 1/8 ━━━━━━━━━━━━━━━━━━━━ 2:04 18s/step - accuracy: 0.3750 - loss: 1.1153 2/8 ━━━━━━━━━━━━━━━━━━━━ 32s 5s/step - accuracy: 0.3516 - loss: 1.2516 3/8 ━━━━━━━━━━━━━━━━━━━━ 28s 6s/step - accuracy: 0.3594 - loss: 1.2565 4/8 ━━━━━━━━━━━━━━━━━━━━ 22s 6s/step - accuracy: 0.3574 - loss: 1.2527 5/8 ━━━━━━━━━━━━━━━━━━━━ 17s 6s/step - accuracy: 0.3584 - loss: 1.2416 6/8 ━━━━━━━━━━━━━━━━━━━━ 11s 6s/step - accuracy: 0.3586 - loss: 1.2344 7/8 ━━━━━━━━━━━━━━━━━━━━ 5s 5s/step - accuracy: 0.3597 - loss: 1.2285 8/8 ━━━━━━━━━━━━━━━━━━━━ 0s 5s/step - accuracy: 0.3607 - loss: 1.2261 8/8 ━━━━━━━━━━━━━━━━━━━━ 72s 8s/step - accuracy: 0.3675 - loss: 1.2099 - val_accuracy: 0.3404 - val_loss: 1.2243 Epoch 2/10 1/8 ━━━━━━━━━━━━━━━━━━━━ 47s 7s/step - accuracy: 0.4062 - loss: 1.1015 2/8 ━━━━━━━━━━━━━━━━━━━━ 41s 7s/step - accuracy: 0.4062 - loss: 1.1044 3/8 ━━━━━━━━━━━━━━━━━━━━ 32s 6s/step - accuracy: 0.3993 - loss: 1.1072 4/8 ━━━━━━━━━━━━━━━━━━━━ 19s 5s/step - accuracy: 0.3915 - loss: 1.1089 5/8 ━━━━━━━━━━━━━━━━━━━━ 16s 6s/step - accuracy: 0.3871 - loss: 1.1089 6/8 ━━━━━━━━━━━━━━━━━━━━ 11s 6s/step - accuracy: 0.3804 - loss: 1.1115 7/8 ━━━━━━━━━━━━━━━━━━━━ 6s 6s/step - accuracy: 0.3777 - loss: 1.1134 8/8 ━━━━━━━━━━━━━━━━━━━━ 0s 6s/step - accuracy: 0.3754 - loss: 1.1151 8/8 ━━━━━━━━━━━━━━━━━━━━ 61s 8s/step - accuracy: 0.3590 - loss: 1.1268 - val_accuracy: 0.3191 - val_loss: 1.1526 Epoch 3/10 1/8 ━━━━━━━━━━━━━━━━━━━━ 46s 7s/step - accuracy: 0.1562 - loss: 1.2461 2/8 ━━━━━━━━━━━━━━━━━━━━ 1:03 11s/step - accuracy: 0.1953 - loss: 1.2096 3/8 ━━━━━━━━━━━━━━━━━━━━ 43s 9s/step - accuracy: 0.2170 - loss: 1.1905 4/8 ━━━━━━━━━━━━━━━━━━━━ 25s 6s/step - accuracy: 0.2241 - loss: 1.1813 5/8 ━━━━━━━━━━━━━━━━━━━━ 20s 7s/step - accuracy: 0.2314 - loss: 1.1736 6/8 ━━━━━━━━━━━━━━━━━━━━ 13s 7s/step - accuracy: 0.2380 - loss: 1.1676 7/8 ━━━━━━━━━━━━━━━━━━━━ 6s 7s/step - accuracy: 0.2429 - loss: 1.1628 8/8 ━━━━━━━━━━━━━━━━━━━━ 0s 7s/step - accuracy: 0.2488 - loss: 1.1585 8/8 ━━━━━━━━━━━━━━━━━━━━ 66s 8s/step - accuracy: 0.2906 - loss: 1.1284 - val_accuracy: 0.3404 - val_loss: 1.0979 Epoch 4/10 1/8 ━━━━━━━━━━━━━━━━━━━━ 49s 7s/step - accuracy: 0.3125 - loss: 1.1056 2/8 ━━━━━━━━━━━━━━━━━━━━ 40s 7s/step - accuracy: 0.3359 - loss: 1.1065 3/8 ━━━━━━━━━━━━━━━━━━━━ 22s 5s/step - accuracy: 0.3366 - loss: 1.1077 4/8 ━━━━━━━━━━━━━━━━━━━━ 22s 6s/step - accuracy: 0.3326 - loss: 1.1093 5/8 ━━━━━━━━━━━━━━━━━━━━ 17s 6s/step - accuracy: 0.3299 - loss: 1.1105 6/8 ━━━━━━━━━━━━━━━━━━━━ 12s 6s/step - accuracy: 0.3249 - loss: 1.1113 7/8 ━━━━━━━━━━━━━━━━━━━━ 6s 6s/step - accuracy: 0.3237 - loss: 1.1113 8/8 ━━━━━━━━━━━━━━━━━━━━ 0s 7s/step - accuracy: 0.3255 - loss: 1.1106 8/8 ━━━━━━━━━━━━━━━━━━━━ 63s 8s/step - accuracy: 0.3376 - loss: 1.1061 - val_accuracy: 0.3404 - val_loss: 1.1047 Epoch 5/10 1/8 ━━━━━━━━━━━━━━━━━━━━ 47s 7s/step - accuracy: 0.3438 - loss: 1.0920 2/8 ━━━━━━━━━━━━━━━━━━━━ 43s 7s/step - accuracy: 0.3359 - loss: 1.1061 3/8 ━━━━━━━━━━━━━━━━━━━━ 36s 7s/step - accuracy: 0.3385 - loss: 1.1070 4/8 ━━━━━━━━━━━━━━━━━━━━ 28s 7s/step - accuracy: 0.3340 - loss: 1.1105 5/8 ━━━━━━━━━━━━━━━━━━━━ 21s 7s/step - accuracy: 0.3284 - loss: 1.1123 6/8 ━━━━━━━━━━━━━━━━━━━━ 14s 7s/step - accuracy: 0.3258 - loss: 1.1136 7/8 ━━━━━━━━━━━━━━━━━━━━ 7s 7s/step - accuracy: 0.3232 - loss: 1.1148 8/8 ━━━━━━━━━━━━━━━━━━━━ 0s 6s/step - accuracy: 0.3213 - loss: 1.1155 8/8 ━━━━━━━━━━━━━━━━━━━━ 62s 8s/step - accuracy: 0.3077 - loss: 1.1205 - val_accuracy: 0.3191 - val_loss: 1.0998 Epoch 6/10 1/8 ━━━━━━━━━━━━━━━━━━━━ 1:11 10s/step - accuracy: 0.3750 - loss: 1.0937 2/8 ━━━━━━━━━━━━━━━━━━━━ 45s 8s/step - accuracy: 0.3672 - loss: 1.0958 3/8 ━━━━━━━━━━━━━━━━━━━━ 37s 7s/step - accuracy: 0.3628 - loss: 1.0970 4/8 ━━━━━━━━━━━━━━━━━━━━ 28s 7s/step - accuracy: 0.3542 - loss: 1.1025 5/8 ━━━━━━━━━━━━━━━━━━━━ 21s 7s/step - accuracy: 0.3508 - loss: 1.1051 6/8 ━━━━━━━━━━━━━━━━━━━━ 12s 6s/step - accuracy: 0.3482 - loss: 1.1068 7/8 ━━━━━━━━━━━━━━━━━━━━ 6s 6s/step - accuracy: 0.3466 - loss: 1.1083 8/8 ━━━━━━━━━━━━━━━━━━━━ 0s 6s/step - accuracy: 0.3433 - loss: 1.1115 8/8 ━━━━━━━━━━━━━━━━━━━━ 65s 8s/step - accuracy: 0.3205 - loss: 1.1336 - val_accuracy: 0.3404 - val_loss: 1.1292 Epoch 7/10 1/8 ━━━━━━━━━━━━━━━━━━━━ 50s 7s/step - accuracy: 0.3438 - loss: 1.1459 2/8 ━━━━━━━━━━━━━━━━━━━━ 38s 6s/step - accuracy: 0.3281 - loss: 1.1426 3/8 ━━━━━━━━━━━━━━━━━━━━ 33s 7s/step - accuracy: 0.3160 - loss: 1.1429 4/8 ━━━━━━━━━━━━━━━━━━━━ 26s 7s/step - accuracy: 0.3132 - loss: 1.1408 5/8 ━━━━━━━━━━━━━━━━━━━━ 20s 7s/step - accuracy: 0.3143 - loss: 1.1383 6/8 ━━━━━━━━━━━━━━━━━━━━ 11s 6s/step - accuracy: 0.3168 - loss: 1.1360 7/8 ━━━━━━━━━━━━━━━━━━━━ 6s 6s/step - accuracy: 0.3182 - loss: 1.1350 8/8 ━━━━━━━━━━━━━━━━━━━━ 0s 6s/step - accuracy: 0.3185 - loss: 1.1348 8/8 ━━━━━━━━━━━━━━━━━━━━ 62s 8s/step - accuracy: 0.3205 - loss: 1.1338 - val_accuracy: 0.3404 - val_loss: 1.1134 Epoch 8/10 1/8 ━━━━━━━━━━━━━━━━━━━━ 48s 7s/step - accuracy: 0.2188 - loss: 1.1531 2/8 ━━━━━━━━━━━━━━━━━━━━ 40s 7s/step - accuracy: 0.2734 - loss: 1.1390 3/8 ━━━━━━━━━━━━━━━━━━━━ 32s 7s/step - accuracy: 0.3038 - loss: 1.1301 4/8 ━━━━━━━━━━━━━━━━━━━━ 25s 6s/step - accuracy: 0.3158 - loss: 1.1257 5/8 ━━━━━━━━━━━━━━━━━━━━ 19s 7s/step - accuracy: 0.3276 - loss: 1.1217 6/8 ━━━━━━━━━━━━━━━━━━━━ 13s 7s/step - accuracy: 0.3303 - loss: 1.1232 7/8 ━━━━━━━━━━━━━━━━━━━━ 6s 6s/step - accuracy: 0.3312 - loss: 1.1243 8/8 ━━━━━━━━━━━━━━━━━━━━ 0s 6s/step - accuracy: 0.3304 - loss: 1.1264 8/8 ━━━━━━━━━━━━━━━━━━━━ 61s 8s/step - accuracy: 0.3248 - loss: 1.1408 - val_accuracy: 0.3191 - val_loss: 1.1156 Epoch 9/10 1/8 ━━━━━━━━━━━━━━━━━━━━ 54s 8s/step - accuracy: 0.3438 - loss: 1.1082 2/8 ━━━━━━━━━━━━━━━━━━━━ 40s 7s/step - accuracy: 0.3203 - loss: 1.1158 3/8 ━━━━━━━━━━━━━━━━━━━━ 33s 7s/step - accuracy: 0.3281 - loss: 1.1146 4/8 ━━━━━━━━━━━━━━━━━━━━ 26s 7s/step - accuracy: 0.3281 - loss: 1.1145 5/8 ━━━━━━━━━━━━━━━━━━━━ 20s 7s/step - accuracy: 0.3325 - loss: 1.1140 6/8 ━━━━━━━━━━━━━━━━━━━━ 13s 7s/step - accuracy: 0.3344 - loss: 1.1162 7/8 ━━━━━━━━━━━━━━━━━━━━ 5s 6s/step - accuracy: 0.3340 - loss: 1.1184 8/8 ━━━━━━━━━━━━━━━━━━━━ 0s 6s/step - accuracy: 0.3344 - loss: 1.1194 8/8 ━━━━━━━━━━━━━━━━━━━━ 61s 8s/step - accuracy: 0.3376 - loss: 1.1264 - val_accuracy: 0.3404 - val_loss: 1.1013 Epoch 10/10 1/8 ━━━━━━━━━━━━━━━━━━━━ 47s 7s/step - accuracy: 0.2812 - loss: 1.1092 2/8 ━━━━━━━━━━━━━━━━━━━━ 40s 7s/step - accuracy: 0.3047 - loss: 1.1057 3/8 ━━━━━━━━━━━━━━━━━━━━ 35s 7s/step - accuracy: 0.3108 - loss: 1.1055 4/8 ━━━━━━━━━━━━━━━━━━━━ 28s 7s/step - accuracy: 0.3112 - loss: 1.1067 5/8 ━━━━━━━━━━━━━━━━━━━━ 21s 7s/step - accuracy: 0.3177 - loss: 1.1066 6/8 ━━━━━━━━━━━━━━━━━━━━ 12s 6s/step - accuracy: 0.3226 - loss: 1.1066 7/8 ━━━━━━━━━━━━━━━━━━━━ 6s 6s/step - accuracy: 0.3253 - loss: 1.1074 8/8 ━━━━━━━━━━━━━━━━━━━━ 0s 6s/step - accuracy: 0.3252 - loss: 1.1092 8/8 ━━━━━━━━━━━━━━━━━━━━ 60s 8s/step - accuracy: 0.3248 - loss: 1.1216 - val_accuracy: 0.3404 - val_loss: 1.1051
mobile = MobileNetV2(
include_top=False,
input_shape=IMG_SIZE+(3,),
weights="imagenet"
)
mobile.trainable = False
model_m = models.Sequential([
mobile,
layers.GlobalAveragePooling2D(),
layers.Dense(256, activation="relu"),
layers.Dense(num_classes, activation="softmax")
])
model_m.compile(
optimizer="adam",
loss="categorical_crossentropy",
metrics=["accuracy"]
)
history_m = model_m.fit(train_ds, validation_data=val_ds, epochs=10)Epoch 1/10 1/8 ━━━━━━━━━━━━━━━━━━━━ 1:24 12s/step - accuracy: 0.4375 - loss: 1.0273 2/8 ━━━━━━━━━━━━━━━━━━━━ 9s 2s/step - accuracy: 0.4092 - loss: 1.1713 3/8 ━━━━━━━━━━━━━━━━━━━━ 17s 3s/step - accuracy: 0.3854 - loss: 1.4973 4/8 ━━━━━━━━━━━━━━━━━━━━ 16s 4s/step - accuracy: 0.3787 - loss: 1.6416 5/8 ━━━━━━━━━━━━━━━━━━━━ 13s 4s/step - accuracy: 0.3740 - loss: 1.6939 6/8 ━━━━━━━━━━━━━━━━━━━━ 9s 5s/step - accuracy: 0.3754 - loss: 1.7062 7/8 ━━━━━━━━━━━━━━━━━━━━ 4s 5s/step - accuracy: 0.3776 - loss: 1.7222 8/8 ━━━━━━━━━━━━━━━━━━━━ 0s 5s/step - accuracy: 0.3806 - loss: 1.7425 8/8 ━━━━━━━━━━━━━━━━━━━━ 69s 8s/step - accuracy: 0.4017 - loss: 1.8848 - val_accuracy: 0.4681 - val_loss: 1.6960 Epoch 2/10 1/8 ━━━━━━━━━━━━━━━━━━━━ 50s 7s/step - accuracy: 0.5312 - loss: 1.4895 2/8 ━━━━━━━━━━━━━━━━━━━━ 37s 6s/step - accuracy: 0.5547 - loss: 1.3683 3/8 ━━━━━━━━━━━━━━━━━━━━ 26s 5s/step - accuracy: 0.5608 - loss: 1.2806 4/8 ━━━━━━━━━━━━━━━━━━━━ 21s 5s/step - accuracy: 0.5768 - loss: 1.2058 5/8 ━━━━━━━━━━━━━━━━━━━━ 16s 6s/step - accuracy: 0.5827 - loss: 1.1586 6/8 ━━━━━━━━━━━━━━━━━━━━ 11s 6s/step - accuracy: 0.5837 - loss: 1.1282 7/8 ━━━━━━━━━━━━━━━━━━━━ 5s 6s/step - accuracy: 0.5807 - loss: 1.1115 8/8 ━━━━━━━━━━━━━━━━━━━━ 0s 5s/step - accuracy: 0.5781 - loss: 1.0971 8/8 ━━━━━━━━━━━━━━━━━━━━ 53s 6s/step - accuracy: 0.5598 - loss: 0.9962 - val_accuracy: 0.3404 - val_loss: 1.3209 Epoch 3/10 1/8 ━━━━━━━━━━━━━━━━━━━━ 45s 6s/step - accuracy: 0.7812 - loss: 0.5883 2/8 ━━━━━━━━━━━━━━━━━━━━ 37s 6s/step - accuracy: 0.7500 - loss: 0.6292 3/8 ━━━━━━━━━━━━━━━━━━━━ 29s 6s/step - accuracy: 0.7431 - loss: 0.6361 4/8 ━━━━━━━━━━━━━━━━━━━━ 23s 6s/step - accuracy: 0.7331 - loss: 0.6470 5/8 ━━━━━━━━━━━━━━━━━━━━ 14s 5s/step - accuracy: 0.7256 - loss: 0.6535 6/8 ━━━━━━━━━━━━━━━━━━━━ 10s 5s/step - accuracy: 0.7194 - loss: 0.6566 7/8 ━━━━━━━━━━━━━━━━━━━━ 5s 5s/step - accuracy: 0.7163 - loss: 0.6576 8/8 ━━━━━━━━━━━━━━━━━━━━ 0s 5s/step - accuracy: 0.7138 - loss: 0.6582 8/8 ━━━━━━━━━━━━━━━━━━━━ 55s 7s/step - accuracy: 0.6966 - loss: 0.6625 - val_accuracy: 0.5319 - val_loss: 1.1151 Epoch 4/10 1/8 ━━━━━━━━━━━━━━━━━━━━ 41s 6s/step - accuracy: 0.8125 - loss: 0.6181 2/8 ━━━━━━━━━━━━━━━━━━━━ 33s 6s/step - accuracy: 0.8203 - loss: 0.5820 3/8 ━━━━━━━━━━━━━━━━━━━━ 28s 6s/step - accuracy: 0.8247 - loss: 0.5635 4/8 ━━━━━━━━━━━━━━━━━━━━ 22s 6s/step - accuracy: 0.8216 - loss: 0.5552 5/8 ━━━━━━━━━━━━━━━━━━━━ 16s 5s/step - accuracy: 0.8210 - loss: 0.5504 6/8 ━━━━━━━━━━━━━━━━━━━━ 10s 5s/step - accuracy: 0.8153 - loss: 0.5505 7/8 ━━━━━━━━━━━━━━━━━━━━ 4s 5s/step - accuracy: 0.8120 - loss: 0.5499 8/8 ━━━━━━━━━━━━━━━━━━━━ 0s 5s/step - accuracy: 0.8109 - loss: 0.5478 8/8 ━━━━━━━━━━━━━━━━━━━━ 49s 6s/step - accuracy: 0.8034 - loss: 0.5324 - val_accuracy: 0.5532 - val_loss: 1.0963 Epoch 5/10 1/8 ━━━━━━━━━━━━━━━━━━━━ 38s 5s/step - accuracy: 0.9062 - loss: 0.4158 2/8 ━━━━━━━━━━━━━━━━━━━━ 10s 2s/step - accuracy: 0.8936 - loss: 0.4158 3/8 ━━━━━━━━━━━━━━━━━━━━ 18s 4s/step - accuracy: 0.8795 - loss: 0.4247 4/8 ━━━━━━━━━━━━━━━━━━━━ 17s 4s/step - accuracy: 0.8719 - loss: 0.4244 5/8 ━━━━━━━━━━━━━━━━━━━━ 13s 5s/step - accuracy: 0.8656 - loss: 0.4232 6/8 ━━━━━━━━━━━━━━━━━━━━ 9s 5s/step - accuracy: 0.8606 - loss: 0.4241 7/8 ━━━━━━━━━━━━━━━━━━━━ 4s 5s/step - accuracy: 0.8579 - loss: 0.4229 8/8 ━━━━━━━━━━━━━━━━━━━━ 0s 5s/step - accuracy: 0.8569 - loss: 0.4215 8/8 ━━━━━━━━━━━━━━━━━━━━ 50s 6s/step - accuracy: 0.8504 - loss: 0.4116 - val_accuracy: 0.4894 - val_loss: 1.1791 Epoch 6/10 1/8 ━━━━━━━━━━━━━━━━━━━━ 50s 7s/step - accuracy: 0.8125 - loss: 0.4535 2/8 ━━━━━━━━━━━━━━━━━━━━ 34s 6s/step - accuracy: 0.8203 - loss: 0.4458 3/8 ━━━━━━━━━━━━━━━━━━━━ 18s 4s/step - accuracy: 0.8262 - loss: 0.4361 4/8 ━━━━━━━━━━━━━━━━━━━━ 17s 4s/step - accuracy: 0.8413 - loss: 0.4171 5/8 ━━━━━━━━━━━━━━━━━━━━ 13s 5s/step - accuracy: 0.8528 - loss: 0.4024 6/8 ━━━━━━━━━━━━━━━━━━━━ 9s 5s/step - accuracy: 0.8626 - loss: 0.3913 7/8 ━━━━━━━━━━━━━━━━━━━━ 4s 5s/step - accuracy: 0.8709 - loss: 0.3814 8/8 ━━━━━━━━━━━━━━━━━━━━ 0s 5s/step - accuracy: 0.8769 - loss: 0.3739 8/8 ━━━━━━━━━━━━━━━━━━━━ 50s 6s/step - accuracy: 0.9188 - loss: 0.3215 - val_accuracy: 0.5319 - val_loss: 1.2046 Epoch 7/10 1/8 ━━━━━━━━━━━━━━━━━━━━ 13s 2s/step - accuracy: 1.0000 - loss: 0.3135 2/8 ━━━━━━━━━━━━━━━━━━━━ 32s 5s/step - accuracy: 0.9762 - loss: 0.2902 3/8 ━━━━━━━━━━━━━━━━━━━━ 26s 5s/step - accuracy: 0.9751 - loss: 0.2836 4/8 ━━━━━━━━━━━━━━━━━━━━ 21s 5s/step - accuracy: 0.9743 - loss: 0.2775 5/8 ━━━━━━━━━━━━━━━━━━━━ 16s 5s/step - accuracy: 0.9722 - loss: 0.2767 6/8 ━━━━━━━━━━━━━━━━━━━━ 10s 5s/step - accuracy: 0.9719 - loss: 0.2723 7/8 ━━━━━━━━━━━━━━━━━━━━ 5s 5s/step - accuracy: 0.9710 - loss: 0.2695 8/8 ━━━━━━━━━━━━━━━━━━━━ 0s 5s/step - accuracy: 0.9698 - loss: 0.2672 8/8 ━━━━━━━━━━━━━━━━━━━━ 48s 7s/step - accuracy: 0.9615 - loss: 0.2508 - val_accuracy: 0.5106 - val_loss: 1.2346 Epoch 8/10 1/8 ━━━━━━━━━━━━━━━━━━━━ 12s 2s/step - accuracy: 1.0000 - loss: 0.2846 2/8 ━━━━━━━━━━━━━━━━━━━━ 33s 6s/step - accuracy: 1.0000 - loss: 0.2407 3/8 ━━━━━━━━━━━━━━━━━━━━ 27s 6s/step - accuracy: 1.0000 - loss: 0.2192 4/8 ━━━━━━━━━━━━━━━━━━━━ 21s 5s/step - accuracy: 0.9976 - loss: 0.2104 5/8 ━━━━━━━━━━━━━━━━━━━━ 16s 5s/step - accuracy: 0.9952 - loss: 0.2046 6/8 ━━━━━━━━━━━━━━━━━━━━ 10s 5s/step - accuracy: 0.9921 - loss: 0.2017 7/8 ━━━━━━━━━━━━━━━━━━━━ 5s 5s/step - accuracy: 0.9897 - loss: 0.1996 8/8 ━━━━━━━━━━━━━━━━━━━━ 0s 5s/step - accuracy: 0.9878 - loss: 0.1978 8/8 ━━━━━━━━━━━━━━━━━━━━ 48s 7s/step - accuracy: 0.9744 - loss: 0.1853 - val_accuracy: 0.5106 - val_loss: 1.3183 Epoch 9/10 1/8 ━━━━━━━━━━━━━━━━━━━━ 40s 6s/step - accuracy: 1.0000 - loss: 0.1067 2/8 ━━━━━━━━━━━━━━━━━━━━ 9s 2s/step - accuracy: 1.0000 - loss: 0.1049 3/8 ━━━━━━━━━━━━━━━━━━━━ 18s 4s/step - accuracy: 0.9955 - loss: 0.1126 4/8 ━━━━━━━━━━━━━━━━━━━━ 17s 4s/step - accuracy: 0.9919 - loss: 0.1183 5/8 ━━━━━━━━━━━━━━━━━━━━ 14s 5s/step - accuracy: 0.9892 - loss: 0.1226 6/8 ━━━━━━━━━━━━━━━━━━━━ 9s 5s/step - accuracy: 0.9880 - loss: 0.1249 7/8 ━━━━━━━━━━━━━━━━━━━━ 4s 5s/step - accuracy: 0.9876 - loss: 0.1270 8/8 ━━━━━━━━━━━━━━━━━━━━ 0s 5s/step - accuracy: 0.9876 - loss: 0.1282 8/8 ━━━━━━━━━━━━━━━━━━━━ 49s 6s/step - accuracy: 0.9872 - loss: 0.1369 - val_accuracy: 0.5106 - val_loss: 1.3777 Epoch 10/10 1/8 ━━━━━━━━━━━━━━━━━━━━ 13s 2s/step - accuracy: 1.0000 - loss: 0.1205 2/8 ━━━━━━━━━━━━━━━━━━━━ 34s 6s/step - accuracy: 1.0000 - loss: 0.1000 3/8 ━━━━━━━━━━━━━━━━━━━━ 27s 6s/step - accuracy: 1.0000 - loss: 0.0976 4/8 ━━━━━━━━━━━━━━━━━━━━ 22s 6s/step - accuracy: 1.0000 - loss: 0.0976 5/8 ━━━━━━━━━━━━━━━━━━━━ 16s 6s/step - accuracy: 1.0000 - loss: 0.0995 6/8 ━━━━━━━━━━━━━━━━━━━━ 11s 6s/step - accuracy: 1.0000 - loss: 0.1011 7/8 ━━━━━━━━━━━━━━━━━━━━ 5s 6s/step - accuracy: 1.0000 - loss: 0.1020 8/8 ━━━━━━━━━━━━━━━━━━━━ 0s 6s/step - accuracy: 1.0000 - loss: 0.1030 8/8 ━━━━━━━━━━━━━━━━━━━━ 48s 7s/step - accuracy: 1.0000 - loss: 0.1102 - val_accuracy: 0.5106 - val_loss: 1.4275
plt.plot(history_e.history['accuracy'], label="EfficientNet Acc")
plt.plot(history_m.history['accuracy'], label="MobileNet Acc")
plt.legend(); plt.title("Comparación Accuracy"); plt.show()
from tensorflow.keras.preprocessing import image
import numpy as np
import matplotlib.pyplot as plt
def predict(img_path, model):
img = image.load_img(img_path, target_size=IMG_SIZE)
img_array = image.img_to_array(img) / 255.0
img_array = np.expand_dims(img_array, 0)
pred = model.predict(img_array)[0]
classes = list(train_ds.class_indices.keys())
result = classes[np.argmax(pred)]
# 🔹 Mostrar la imagen y la predicción
plt.imshow(image.load_img(img_path))
plt.title(f"Predicción: {result}")
plt.axis("off")
plt.show()
return result, pred
img_path = r"G:\Mi unidad\Colab Notebooks\AProfundo\clasificacion_dataset\test\monilia\Monilia101.jpg"
print(predict(img_path, model_m))1/1 ━━━━━━━━━━━━━━━━━━━━ 0s 2s/step 1/1 ━━━━━━━━━━━━━━━━━━━━ 2s 2s/step
('monilia', array([2.3191734e-01, 7.6796454e-01, 1.1810561e-04], dtype=float32))